> How about using a RESON filter, with the bandwidth selected as a very
> small fraction of the cutoff frequency (for very high Q), and normalized
> to 1? It won't self-oscillate, but it will ring for a good long time
> when excited by a signal.

I will try. For the moment, I'm using moogvcf with high Q settings
as suggested by Mike Berry. It works quite well.
> 
> As far as converting the moogvcf to a true bandpass filter, I don't
> think it can be done. In the analog realm, cascade-style filters were
> never used for bandpass - a state-variable filter was the filter
> normally used (Csound might have a Chamberlin style state-variable
> filter; if not, I might try and code one soon).  A Moog highpass would

A state-variable-filter - oh yes yes, please, could you code one ?
I'd love to have another excellent sounding filter like 'moogvcf'.
And as far as I know, the state variable filter can be used as
lowpass, bandpass, highpass and notch (band-reject) at the same time,
depending on where you take the output (in the analog circuit, I have
no indea how it is in the digital implementation). So we would have
another cool filterset, let's say 'statelp', 'statebp', 'statehp',
'statebr'. I'm looking forward for it. In the meantime, I work with
what's already there, and that's quite a lot. Wait for my second song.
Already now it is _much_ better than the first (which I called second (-;),
and it's not even finished. This will become a real song you could play
at a party, not just a first test like the other one I put on my ftp
site recently...


> Does the moogvcf opcode self-oscillate? I am under the impression that
> no digital filter self-oscillates without an initial source of
> excitation, but I could be wrong.

It does self-oscillate, yes. I think that the moogvcf is not a classical
digital filter, it's a simulation of an analog circuit filter